Vol. 9, No. 4 , April 1970
THEFe(1II)-EDTA-H2O2 SYSTEM931 CONTRIBUTION FROM THE DEPARTMENT OF CHEMISTRY, COLUMBIA UNIVERSITY, NEW YORK,NEW YORK 10027
The Iron(II1)-Ethylenediaminetetraacetic Acid-Peroxide System] BY CHEVES WALLING,2 MICHAEL KURZ,
AND
HARVEY J. SCHUGAR
Received Decenzbev 22, 1969 The purple complex formed in alkaline solutions containing Fe(III), EDTA, and HZOzhas been shown by spectrophotometric, potentiometric, and magnetic measurements to have the stoichiometric composition Fe111(EDTA)02a-. The complex is a catalyst for the decomposition of HzOz. At p H 2 10.5 decomposition is cleanly to HzO and Oz. The rate shows a maximum near p H 9, and in this region a variety of organic substrates (including EDTA and the complex itself) are oxidized as well, the overall rate of HzOz decomposition varying greatly with the substrates present. Evidence for the evolution of singlet oxygen in these systems was negative. Some analogs of EDTA behave similarly as catalysts, nitrilotriacetic acid being particularly effective
In 1956 Cheng and Lott3 reported the formation of a purple color by the interaction of excess hydrogen peroxide with the FeI'I-EDTA (ethylenediaminetetraacetic acid) complex in alkaline solution and offered two possible explanations for the phenomenon: a peroxy complex of FeI'IEDTA or chelated iron in a higher oxidation state. Both hypotheses have had subsequent adherents. Ringbom and coworkers4 have interpreted their spectral evidence as supporting a peroxy complex formed from FeIIIEDTA and HOO-, and others5r6 have reached similar conclusions. Recently the rate of color formation has been investigated by stop-flow methods and was reported to obey the expression rate = k[EDTAFe(OH)z3-] [HzOZ]
(1)
over the pH range 9.0-11.0.7 On the other hand a complexed ferrate ion structure has also been proposed for the colored substance, also on the basis of spectral evidence.8 To date the purple complex is only known in solution, and is, in fact, quite unstable. It catalyzes the decomposition of H20z,9,10 and the system has also been reported t o initiate the emulsion polymerization of styrene.1° Because of our interest in metal-catalyzed redox systems involving peroxides we decided to investigate this system. The work k d initially to a study of some other properties of ferric complexes in alkaline solution, 11,12 but we now report spectral, potentiometric, and magnetic susceptibility data which essentially confirm Ringbom's f o r m ~ l a t i o n . ~In addition (1) Support of this work by a grant from the National Science Foundation is gratefully acknowledged. (2) T o whom inquiries should be addressed at the Department of Chemistry, University of Utah, Salt Lake City, Utah 84112. (3) K. L. Cheng and P. F. Lott, Anal. Chem., 28, 462 (1956). (4) A. Ringborn, S. Siitonen, and B. Saxen, Anal. Chim. Acta, 16, 541 (1957). ( 5 ) P. M . Wader, J. Ant. Chenc. SOL.,88, 2956 (1960). (6) Z. P. Kachanova and A. P. Purmal, Russ. J . Phys. Chem., 38, 1342 (1964). (7) M. Orhanovic and R. G. Wilkins, Cvoat. Chim. Acta, 39, 149 (1967). (8) G. L. Kochanny, Jr., and A. Timnick, J . A m . Chem. SOC.,83, 2777 (1961). (9) Z. P . Kachanova and A. P. Purmal, Russ. J . Phys. Chem., 38, 1360 (1964). (10) J. Bond, C. W. Brown, and G. S. Rushton, Polymev Letfeus, 2, 1015 (1964). (11) H. Schugar, C. Walling, R. B. Jones, and Il. B. Gray, J . A m . Chem. SOL.,89, 3712 (1967). (12) S. J. Cippard, H. Schugar, and C. Walling, Inovg. Chem., 6, 1825 (1967).
we hive examined some properties of the decomposition of the complex, both in the presence and in the absence of organic substrates. These reactions turn out t o be surprisingly complex, the system showing both catalase and peroxidase activity, decomposing H202 to oxygen and water and attacking organic substrates, the relative importance of the two paths depending upon pH. Results Spectrophotometric Studies.-On the basis of X-ray crystal structure determinations, Hoard has concluded that in weakly acid solutions the ferric EDTA complex is heptacoordinate and binds 1 equiv of water: Fe(EDTA)H20-.1ss14 A t higher pH's it behaves as a weak acid, 16f16 presumably changing to Fe(EDTA)OH2-, and also undergoes dimerization (olation) l6 to form an oxo-bridged structure, Fe(EDTA)OFe(EDTA)4-.11*12Thus a t least two equilibria are involved with equilibrium constants a t 13"
e
Fe(EDTA)HzOFe(EDTA)OH2-
=F=
+ H+
2Fe(EDTA)OH2Fez(EDTA)z04-
+ HzO
K Z = 1.6 X
(2)
K3 = 4 X lo2 l7 (3)
and a t the pH's with which we are concerned the complex is almost entirely in the hydroxy and dimeric forms, Addition of hydrogen peroxide leads to the formation of a new purple complex, for which we obtain X,, 520 nm, e 528 (compared with 547, 543, and 530 reported earlier), and for which we find the same stoichiometry reported by Ringborn. (Equation 1is also the stoichiometric equivalent of eq 4.) Fe(EDTA)OHZ- $. HOz-
Fe(EDTA)0z3-
+ HzO
(4)
The equilibrium constant for (4) was calculated as described in the Experimental Part, both ignoring the dimerization ( 3 ) which was relatively unimportant in most systems and taking i t into account. Results appear in Table I. The two calculations give com(13) M. 13. Lind, M . J. Hamor, T. A. Hamor, and J. L. Hoard, ibid., 3, 34 (1964). (14) E. L. Muetterties and C. M. Wright, Quavt. Rev. (London), 21, 109 (1967). (15) G. Schwarzenbachand J. Heller, Helv. Chim. Acta, 34, 576 (1951); (16) R. L.bustafson and A. E. Martell, J . Phys. Chem., 6 7 , 576 (1963). (17) H. J. Schugar, A. T. Hubbard, F. C. Anson, and H. B . Gray, J . A m . Chem. SOL.,91, 71 (1969); Gustafson and Martell gave 890.10
Inorganic Chemistry
M. KURZ,AND H . J. SCHUGAR ‘932 C. WALLING, TABLE I SPECTROMETRIC FORMATION CONSTASTS FOR Fe111(EDTA4)02COXPLEX AT 13 pH
lO~[HzO&,bM
lO~[[Fe]t?Ji
Log K’c
p (B. hi. )
5.0
THE
4.0 34.8 4.0 0.39 4.0 4.0 4.0 1.0 4.0 0.39 4.0 4.0 4.0 4.0
2.0 2.0 2.0 2.0 10.0 6.0 2.0 6.0 2.0 2.0 2.0 2.0 2.0 2.0
parable overall consistency, but the second does show a smaller scatter a t low pH and high iron concentration where dimerization becomes appreciable. Both sets also indicate an apparent decrease in K’s a t higher pH’s, perhaps due to further transformations of the Fe(EDTA)OH complex (the complex is reported to ionize further a t high pHI5 and is also somewhat unstable). If we take dimerization into account but neglect data a t pH >10.5, we obtain K = 2.4 X l o 4 with an average deviation of 12’%. This is slightly higher than previous values [(3.9-9.8) X 103]4t537obtained a t 25’. Potentiometric Studies.-At pH’s a t which HzOzis only slightly ionized, eq 4 predicts that addition of Hz02 to the ferric EDTA complex will lead to the liberation of acid. The prediction was investigated by mixing solutions of HzOzand complex, adjusted to the same pH, and determining the amount of base required t o restore the pH to its original value. The amount of base required may be expressed in terms of a pseudoequilibrium constant (eq 5), where [OH-] represents the added base. A typical set of data in which HzOz K“
=
[OH-l/([Fel - [OH-])([H201] - [OH-])
(5)
was added to a ferric EDTA solution, both a t pH 10.5) and the amount of base required to readjust the pH determined appears in Table 11. Consistency TABLE I1 EVALUATION O F EQ 5 HzOP
K ‘I
HzOiQ
K”
HzOP
5.4
5.6
5.8
6.0
I
I
I
I
I
Log K C
4.20 4.47 4.39 4.48 4.24 4.47 4.01 4.29 4.03 4.52 3.97 4.39 4.41 4.24 4.05 4.44 4.23 4.32 4.11 4.31 4.18 4.22 4.23 4.23 4.07 4.05 4.02 3.99 Av 4.14 4.33 Av dev 0.10 0.13 All experiments in the presence of 0.24 M NaC104, 8% tbutyl alcohol, and fivefold excess EDTA. * Total concentrations. c K and K’ were calculated with and without taking dimerization into account. 8.2 8.5 8.5 8.5 8.5 8.5 9.0 9.0 9.5 9.5 10.0 10.5 11.0 11.5
5.2
K”
0.05 6.20 0.20 6.09 0.50 6.00 0.10 6.35 0.30 6.25 0.70 6.08 a Millimoles of HzO2 added to 0.10 mmol of ferric EDTA.
is good, and very poor fits were obtained using other stoichiometries. Magnetic Susceptibility Measurements.-The magnetic moment of the ferric EDTA complex has been
11
pH
10
9
*t I
0
I
.2
I .4
I
I
.6
.8
I
1.u
Absorbance
Figure 1.-Comparison of magnetic moment and absorbance of the Fe(III)-EDTA-H202 complex.
reported as 6.07 BM a t pH 6 (theoretical for high-spin iron, 5.92 BM) but decreases a t higher pH due to formation of the low-spin dimer.ll The magnetic behavior of the peroxy complex was investigated by the nmr method of Evans18 and is shown in Table 111. At pH 8.5 there is some drop TABLE 111 MAGNETIC PROPERTIES OF Fe(EDTA)023- a Splitting, pH
CPS
103Xm,cgsu
fir
BM
Absorbance (520 nm)
6 . Ob 3.96 15.74 5.97 ... 8.5 2.92 11.61 5.13 0.19 9.5 3.65 14.51 5.74 0.82 10.5 4.04 16.06 6.03 1.02 11.5 4.09 16.26 6.07 1.05 a Using 0.042 M H202, 0.002 M Fe, 0.01 M EDTA, and 8% t-butyl alcohol a t 8”. * No HnOz.
in moment due to dimer formation, but values a t pH 10.5 and 11.5, where peroxy complex formation is almost complete, give the moment of high-spin Fe(II1). Plots of magnetic moment and absorbance a t 520 nm us. pH yield similar curves, Figure 1, and it is evident that the results are inconsistent with formulations involving other valence states of iron. All the observations which we have described are consistent with a formulation of the complex formed in Fe(III)-EDTA-H202 systems as Fe111(EDTA)0z3or some stoichiometric equivalent such as Fe”‘(EDTA)02HOH3-. The stoichiometry is inconsistent with a dimeric structure, and magnetic susceptibility measurements and our potentiometric measurements (plus the simple observation that the complex may be repeatedly decomposed and regenerated by lowering and raising pH) rule out any higher oxidation state of iron.lg Of the possible stoichiometric equivalents, Fe111(EDTA)023- appears the most plausible on the basis of several arguments. It is analogous to a number (18) D. F. Evans, J . Chem. S O L .2003 , (1959). (19) If, for example Fe(II1) were oxidized to ferrate, 5 equiv of protons would be liberated compared with 1 equiv observed.
VoZ. 9, No. 4, April 1970
THEFe(III)-EDTA-H202 SYSTEM933
of other known species Fe(EDTA)X2-; similar colored species are formed from a number of other ferric chelates a t high pH (vide infra), but we detect no colored complex from Fe(EDTA)OH2- and either hypochlorite or tbutyl hydroperoxide. I n fact such systems are stable a t room temperature and give no evidence of reaction. Accordingly a second ionizable hydrogen appears to be involved in complex formation. Finally, the optical absorption observed is not simply a bathochromic shift of the EDTA complex absorption, but a new peak and presumably a charge-transfer band, plausibly associated with the electron-transfer process
e FeI1(EDTA)O2*-
Fe1I1(EDTA)OZ2-
TABLE IV EFFECT OF IRONCONCENTRATION ON H 2 0 2 DECOMPOSITION AT pH 10.54 d[Oz]/dt, mmol/min
lO"Fe1, M
d[Otl/dt, mmol/min
2 0.006 6 0.024 4 0.013 8 0.029 a [HzOz] = 0.040 M , 2 mmol total; [EDTA] = 0.0020 M; temperature 30'.
mum decomposition rate, HzO2 both decomposes to O2 and attacks EDTA, oxidizing it to COZ and ammonia. Since this reaction liberates acid, i t is conveniently followed by measuring the amount of sodium hydroxide which must be added to maintain constant pH. Because the reaction destroys the complex, systems not containing excess EDTA soon turn from (20) M . G. Evans, R. George, and N. Uri, Trans. Faraday Soc (1949).
I
Acid
I
-
/
(6)
facilitated by the stability of the superoxide anion. Thus in contrast hydroperoxy complexes of hydrated ferric ion simply show a bathochromic shift of usual ferric ion absorption.20 If this interpretation is correct, the driving force for complex formation can be rationalized as involving the replacement of OH- by more nucleophilic OOH-. The resulting complex should be more acidic than H202,it loses a proton to the alkaline medium, and the resulting 0 2 2 - ion is even more strongly bound to iron. Catalytic Properties of the Fe (III)-EDTA-H202 System.-In alkaline solution FeIII-EDTA complexes catalyze the decomposition of hydrogen peroxide and its oxidation of a variety of organic substrates including free and complexed EDTA. The phenomena involved show complicated pH and concentration dependence and a t this point we can give little more than a descriptive summary of our findings with some brief comments about possible reaction paths. The rate of HzO2 decomposition is very slow below pH 8, rises to a flat maximum a t pH 9-10, and then declines. At pH 10.5 and above decomposition is cleanly to O2 and water; organic substrates are not attacked and have no effect on rate, which, however, decreases slightly with increasing ionic strength. Oxygen evolution is close to zero order over most of the reaction and apparently first order in iron, which exists almost entirely as a peroxy complex under these conditions, Table IV. A t lower pH's in the region of maxi-
104[Fe], M
moles/mole H20,
, 45, 230
Time ( m i n )
Figure 2.-Formation of acids and 0 2 in glycol oxidation: HzOz,0.04 M ; Fe, 0.0004 M; EDTA, 0.002 If; glycol, 0.68 M; pH 9; temperature 30".
violet to yellow and then precipitate ferric hydroxide.21 At this point the rate of HzOz decomposition drops sharply. I n this pH range other substrates compete with EDTA for oxidation by hydrogen peroxide and have large effects on the overall rate of decomposition which can be either accelerated or retarded. Since the effect of ethylene glycol was particularly large, it was studied in relative detail. TABLE V EFFECTOF pH ON THE Fe-EDTA-HnOz-ETHYLENE GLYCOLREACTION^ PH
[Products], mol/mol of HiOz Acids 0 2
Max. rate, mmol/min Acids 0 2
7.75 0.40 0.05 0.024 0,002 8.0 0.38 0.06 0.037 0.004 8.5 0.35 0.10 0.071 0.012 9.0 0.27 0.19 0.054 0,030 9.5 0.17 0.32 0.027 0.052 10.0 0.09 0.39 0.011 0.046 10.5 0 0.48 0 0,015 a [HzOZ]= 0.040 M , 2 mmol total; [Fe] = 0.0004 M; [glycol] = 0.068 M ; [EDTA] = 0.0020 M ; temperature 30".
Oxidation of Ethylene Glycol.-After a short induction period, H 2 0 zdecomposition is markedly accelerated in the presence of ethylene glycol. Both 02 and acid are liberated, and the glycol is attacked in preference to EDTA,22effectively protecting the complex from decomposition even when little excess EDTA is present. Both rate and product distribution are pH dependent, Table V, but material balances calculated from the product data given indicate that nearly 0.50 equiv of acid is produced per molecule of HzOz reacting with the glycol a t all pH's. I n any given experiment both oxygen and acid liberations follow sigmoid curves, and some typical plots appear in Figure 2 . Rates reported in Table rJ and elsewhere are maximum values (21) This phenomeonon was described by Kachanova a n d Purma1,o who however did not recognize its significance. (22) T h e stoichiometry (moles of acid liberated/mole of Hi02 consumed) may vary from zero t o unity for different oxidations. However for oxidation of glycol t o glycolate, oxalate, or bicarbonate i t is 0.5, 0.5, a n d 0.4, respectively (see below).
Inorganic Chemistry
934 C. WALLING, M. KURZ,AND H. J. SCHUGAR
calculated from the linear portions of the curves. These maxima show a complex dependence on concentration of reactants. Excess EDTA retards the reaction, Table VI, further evidence that oxidations of EDTA
EFFECTOF
THE
I
TABLE VI [EDTA]/[Fe] RATIOON GLYCOLOXIDATION^
[EDTAI/[Fel
Acids formed
5 3 2 1 0.5
Max. rate
0.056 0.060
0.30 0.30 0.32 0.35 0.23b
0.088 0.242 0.112 0
